Training in progress, epoch 3, checkpoint

last-checkpoint/global_step396/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:30c5019d1193156a92602470d7ae99ebac786b447e13ea5f8c5115b0dc32d989
+size 13476836524

last-checkpoint/global_step396/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea72d8c70757512ab84c594075385908d297dacf90215a8ff21446fc8a5cc76d
+size 13476836524

last-checkpoint/global_step396/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb156e02fd3af4a02431a8226dc3c45bb5443d3813c8052e7c8786ed2fb9a515
+size 13476836524

last-checkpoint/global_step396/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70698e535be40d0af25693ea5a97e97d728881da921446fe6feeccd02cbebaed
+size 13476836524

last-checkpoint/global_step396/zero_pp_rank_0_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5962b2aca10696a8628b10cbce7ad78a11000b3fac1d95a756963951ceffad3e
+size 150693

last-checkpoint/global_step396/zero_pp_rank_1_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3691154be3e3e35e1d4e0345e8aeacdb622d3637e3091010ddb0619d9d658649
+size 150693

last-checkpoint/global_step396/zero_pp_rank_2_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d4e16b1c445fd3010969c2e4ce5b31c7d615bd0e015a33c0eeb08d188183b55
+size 150693

last-checkpoint/global_step396/zero_pp_rank_3_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6722ba9b3365e1978cc112963745f06b86d81e638b5bbd3b50600f968f5630be
+size 150693

last-checkpoint/latest

@@ -1 +1 @@
-global_step132
+global_step396

last-checkpoint/model-00001-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:db66773ea1122a0d2d3af137fe42ffb319ac54cd207a4692701f78b80c71fed5
+oid sha256:1b71ccd6b4bc4e75b9800c25d2fd338c358d153dc7ee8bf8b616968c3af23eb8
 size 4938985352

last-checkpoint/model-00002-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:374333ea4d29f3944f0646aeb4c677f98e88ad94ee954caff38ccd4994248606
+oid sha256:7c5b6273ae38851bedf366abcc0a02c88104b0c110c69c9388aedf49c0d09c65
 size 4947390880

last-checkpoint/model-00003-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:95ba0b432b979a6cb0ad3cd1780233a076617c586d73ab17d8e46694f738294f
+oid sha256:38b4b1c051ff762cba0f3f8d8b562f578ba9a873bfbda971b0a90deb216d0cf5
 size 3590488816

last-checkpoint/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f1ce770862b76ceecb3453a5d0b9ef65311eab0f6a60f52997726a89d718de08
+oid sha256:ab7cf0d34d60dfda516cf9661904550e2e294e723edd07c25c738f05e8ba92d1
 size 15024

last-checkpoint/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2bcae6573442578b752be3e988d4ccb38056b45c31b5c02eb579ec6cebcfa62e
+oid sha256:f7b183324e8227a51a9556d86b2ad893a8c4c52205ed4a737356c6611dac7353
 size 15024

last-checkpoint/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:398ff45072fee2975e88b9e078f915103d75d94bc08753303cf855b915973623
+oid sha256:ac69e994090f4818cb1fa6f6cefa363178552c3c731c6507ff195bcb07fd5bef
 size 15024

last-checkpoint/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7f4b210bc832d4c648e3bee8c21dca26e5a8b365d6ec90c638062005a052e57b
+oid sha256:f68e61b63402f8afb1f69c960f7944965655dac11e3ccf29919c282f23931f86
 size 15024

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:081681bbbe98e112e0df8f351c9ba3661c598d1ed2ac20b20f26378a5328c1e6
+oid sha256:bd42a1a947062e05f4c7c69eaa1bb1d4fd05329e7438195743e9f0a70c1343ae
 size 1064

last-checkpoint/trainer_state.json

@@ -1,9 +1,9 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.0,
+  "epoch": 3.0,
   "eval_steps": 500,
-  "global_step": 132,
+  "global_step": 396,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -233,6 +233,428 @@
       "eval_samples_per_second": 16.271,
       "eval_steps_per_second": 0.569,
       "step": 132
+    },
+    {
+      "epoch": 1.0606060606060606,
+      "grad_norm": 0.025427232520216565,
+      "learning_rate": 3.5955056179775277e-07,
+      "logits/chosen": -1.5546875,
+      "logits/rejected": -1.421875,
+      "logps/chosen": -156.0,
+      "logps/rejected": -243.0,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.9140625,
+      "rewards/margins": 10.5625,
+      "rewards/rejected": -11.5,
+      "step": 140
+    },
+    {
+      "epoch": 1.1363636363636362,
+      "grad_norm": 0.01529093932750603,
+      "learning_rate": 3.4550561797752807e-07,
+      "logits/chosen": -1.53125,
+      "logits/rejected": -1.46875,
+      "logps/chosen": -140.0,
+      "logps/rejected": -250.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.76171875,
+      "rewards/margins": 11.125,
+      "rewards/rejected": -11.875,
+      "step": 150
+    },
+    {
+      "epoch": 1.2121212121212122,
+      "grad_norm": 0.010074912036860673,
+      "learning_rate": 3.314606741573033e-07,
+      "logits/chosen": -1.5390625,
+      "logits/rejected": -1.4453125,
+      "logps/chosen": -150.0,
+      "logps/rejected": -245.0,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.8984375,
+      "rewards/margins": 11.0,
+      "rewards/rejected": -11.875,
+      "step": 160
+    },
+    {
+      "epoch": 1.2878787878787878,
+      "grad_norm": 0.024954993930167084,
+      "learning_rate": 3.1741573033707863e-07,
+      "logits/chosen": -1.59375,
+      "logits/rejected": -1.4921875,
+      "logps/chosen": -152.0,
+      "logps/rejected": -247.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.76953125,
+      "rewards/margins": 10.875,
+      "rewards/rejected": -11.625,
+      "step": 170
+    },
+    {
+      "epoch": 1.3636363636363638,
+      "grad_norm": 0.003532695770478855,
+      "learning_rate": 3.0337078651685393e-07,
+      "logits/chosen": -1.5078125,
+      "logits/rejected": -1.46875,
+      "logps/chosen": -127.5,
+      "logps/rejected": -249.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.9765625,
+      "rewards/margins": 10.9375,
+      "rewards/rejected": -11.875,
+      "step": 180
+    },
+    {
+      "epoch": 1.4393939393939394,
+      "grad_norm": 0.0018665433852718737,
+      "learning_rate": 2.893258426966292e-07,
+      "logits/chosen": -1.6015625,
+      "logits/rejected": -1.5078125,
+      "logps/chosen": -164.0,
+      "logps/rejected": -258.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.91015625,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -12.25,
+      "step": 190
+    },
+    {
+      "epoch": 1.5151515151515151,
+      "grad_norm": 0.020725382536818583,
+      "learning_rate": 2.752808988764045e-07,
+      "logits/chosen": -1.578125,
+      "logits/rejected": -1.4765625,
+      "logps/chosen": -157.0,
+      "logps/rejected": -253.0,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.0078125,
+      "rewards/margins": 11.25,
+      "rewards/rejected": -12.3125,
+      "step": 200
+    },
+    {
+      "epoch": 1.5909090909090908,
+      "grad_norm": 0.009333169079013727,
+      "learning_rate": 2.612359550561798e-07,
+      "logits/chosen": -1.515625,
+      "logits/rejected": -1.4765625,
+      "logps/chosen": -154.0,
+      "logps/rejected": -260.0,
+      "loss": 0.0005,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.390625,
+      "rewards/margins": 11.4375,
+      "rewards/rejected": -12.875,
+      "step": 210
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 0.018292778581923616,
+      "learning_rate": 2.4719101123595505e-07,
+      "logits/chosen": -1.5703125,
+      "logits/rejected": -1.375,
+      "logps/chosen": -156.0,
+      "logps/rejected": -248.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.2265625,
+      "rewards/margins": 11.375,
+      "rewards/rejected": -12.625,
+      "step": 220
+    },
+    {
+      "epoch": 1.7424242424242424,
+      "grad_norm": 0.006108685272396747,
+      "learning_rate": 2.331460674157303e-07,
+      "logits/chosen": -1.5625,
+      "logits/rejected": -1.40625,
+      "logps/chosen": -155.0,
+      "logps/rejected": -253.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.2578125,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -12.75,
+      "step": 230
+    },
+    {
+      "epoch": 1.8181818181818183,
+      "grad_norm": 0.008540878901725096,
+      "learning_rate": 2.191011235955056e-07,
+      "logits/chosen": -1.5546875,
+      "logits/rejected": -1.4921875,
+      "logps/chosen": -145.0,
+      "logps/rejected": -256.0,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.046875,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -12.875,
+      "step": 240
+    },
+    {
+      "epoch": 1.893939393939394,
+      "grad_norm": 0.26338726379953975,
+      "learning_rate": 2.0505617977528089e-07,
+      "logits/chosen": -1.5625,
+      "logits/rejected": -1.5,
+      "logps/chosen": -152.0,
+      "logps/rejected": -255.0,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.046875,
+      "rewards/margins": 11.8125,
+      "rewards/rejected": -12.875,
+      "step": 250
+    },
+    {
+      "epoch": 1.9696969696969697,
+      "grad_norm": 0.0326586506464961,
+      "learning_rate": 1.9101123595505617e-07,
+      "logits/chosen": -1.5703125,
+      "logits/rejected": -1.4453125,
+      "logps/chosen": -169.0,
+      "logps/rejected": -258.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1875,
+      "rewards/margins": 11.625,
+      "rewards/rejected": -12.8125,
+      "step": 260
+    },
+    {
+      "epoch": 2.0,
+      "eval_logits/chosen": -1.5,
+      "eval_logits/rejected": -1.484375,
+      "eval_logps/chosen": -149.0,
+      "eval_logps/rejected": -260.0,
+      "eval_loss": 6.90150263835676e-05,
+      "eval_rewards/accuracies": 1.0,
+      "eval_rewards/chosen": -1.0078125,
+      "eval_rewards/margins": 11.8125,
+      "eval_rewards/rejected": -12.8125,
+      "eval_runtime": 14.6298,
+      "eval_samples_per_second": 13.671,
+      "eval_steps_per_second": 0.478,
+      "step": 264
+    },
+    {
+      "epoch": 2.0454545454545454,
+      "grad_norm": 0.0013074998122167467,
+      "learning_rate": 1.7696629213483144e-07,
+      "logits/chosen": -1.546875,
+      "logits/rejected": -1.4609375,
+      "logps/chosen": -155.0,
+      "logps/rejected": -258.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.90625,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.8125,
+      "step": 270
+    },
+    {
+      "epoch": 2.121212121212121,
+      "grad_norm": 0.018655034734646705,
+      "learning_rate": 1.6292134831460675e-07,
+      "logits/chosen": -1.59375,
+      "logits/rejected": -1.46875,
+      "logps/chosen": -166.0,
+      "logps/rejected": -258.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.0546875,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -13.25,
+      "step": 280
+    },
+    {
+      "epoch": 2.196969696969697,
+      "grad_norm": 0.014607729180860774,
+      "learning_rate": 1.4887640449438203e-07,
+      "logits/chosen": -1.59375,
+      "logits/rejected": -1.484375,
+      "logps/chosen": -170.0,
+      "logps/rejected": -262.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.96875,
+      "rewards/margins": 11.9375,
+      "rewards/rejected": -12.9375,
+      "step": 290
+    },
+    {
+      "epoch": 2.2727272727272725,
+      "grad_norm": 0.009493204077863557,
+      "learning_rate": 1.3483146067415728e-07,
+      "logits/chosen": -1.609375,
+      "logits/rejected": -1.546875,
+      "logps/chosen": -160.0,
+      "logps/rejected": -260.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.9921875,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -13.125,
+      "step": 300
+    },
+    {
+      "epoch": 2.3484848484848486,
+      "grad_norm": 0.003736203163504444,
+      "learning_rate": 1.2078651685393259e-07,
+      "logits/chosen": -1.578125,
+      "logits/rejected": -1.5390625,
+      "logps/chosen": -152.0,
+      "logps/rejected": -268.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.0234375,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -13.25,
+      "step": 310
+    },
+    {
+      "epoch": 2.4242424242424243,
+      "grad_norm": 0.0053623346535288585,
+      "learning_rate": 1.0674157303370785e-07,
+      "logits/chosen": -1.5546875,
+      "logits/rejected": -1.4453125,
+      "logps/chosen": -157.0,
+      "logps/rejected": -260.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1015625,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -13.125,
+      "step": 320
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 0.001064339902935614,
+      "learning_rate": 9.269662921348314e-08,
+      "logits/chosen": -1.546875,
+      "logits/rejected": -1.515625,
+      "logps/chosen": -164.0,
+      "logps/rejected": -260.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1953125,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -13.375,
+      "step": 330
+    },
+    {
+      "epoch": 2.5757575757575757,
+      "grad_norm": 0.04406263550084857,
+      "learning_rate": 7.865168539325842e-08,
+      "logits/chosen": -1.5390625,
+      "logits/rejected": -1.4921875,
+      "logps/chosen": -164.0,
+      "logps/rejected": -264.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.95703125,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.625,
+      "step": 340
+    },
+    {
+      "epoch": 2.6515151515151514,
+      "grad_norm": 0.002620164881663117,
+      "learning_rate": 6.460674157303371e-08,
+      "logits/chosen": -1.5078125,
+      "logits/rejected": -1.5078125,
+      "logps/chosen": -144.0,
+      "logps/rejected": -272.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.0390625,
+      "rewards/margins": 12.5,
+      "rewards/rejected": -13.5,
+      "step": 350
+    },
+    {
+      "epoch": 2.7272727272727275,
+      "grad_norm": 0.016885587735988555,
+      "learning_rate": 5.056179775280899e-08,
+      "logits/chosen": -1.5625,
+      "logits/rejected": -1.4765625,
+      "logps/chosen": -160.0,
+      "logps/rejected": -264.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1484375,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -13.3125,
+      "step": 360
+    },
+    {
+      "epoch": 2.8030303030303028,
+      "grad_norm": 0.01685236462277552,
+      "learning_rate": 3.6516853932584266e-08,
+      "logits/chosen": -1.5078125,
+      "logits/rejected": -1.453125,
+      "logps/chosen": -137.0,
+      "logps/rejected": -262.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1015625,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -13.5,
+      "step": 370
+    },
+    {
+      "epoch": 2.878787878787879,
+      "grad_norm": 0.0035252409898577407,
+      "learning_rate": 2.2471910112359548e-08,
+      "logits/chosen": -1.515625,
+      "logits/rejected": -1.4609375,
+      "logps/chosen": -142.0,
+      "logps/rejected": -268.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.078125,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -13.6875,
+      "step": 380
+    },
+    {
+      "epoch": 2.9545454545454546,
+      "grad_norm": 0.002976821998771009,
+      "learning_rate": 8.42696629213483e-09,
+      "logits/chosen": -1.546875,
+      "logits/rejected": -1.515625,
+      "logps/chosen": -158.0,
+      "logps/rejected": -268.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1484375,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -13.6875,
+      "step": 390
+    },
+    {
+      "epoch": 3.0,
+      "eval_logits/chosen": -1.5,
+      "eval_logits/rejected": -1.4921875,
+      "eval_logps/chosen": -149.0,
+      "eval_logps/rejected": -264.0,
+      "eval_loss": 6.22963925707154e-05,
+      "eval_rewards/accuracies": 1.0,
+      "eval_rewards/chosen": -1.046875,
+      "eval_rewards/margins": 12.1875,
+      "eval_rewards/rejected": -13.25,
+      "eval_runtime": 14.9995,
+      "eval_samples_per_second": 13.334,
+      "eval_steps_per_second": 0.467,
+      "step": 396
     }
   ],
   "logging_steps": 10,
@@ -247,7 +669,7 @@
         "should_evaluate": false,
         "should_log": false,
         "should_save": true,
-        "should_training_stop": false
+        "should_training_stop": true
       },
       "attributes": {}
     }